Automatic call routing (ACR) systems are commonplace for businesses that process high call volumes. A goal of an ACR system is to determine why a customer is calling and to route the customer to an appropriate service department. In order for an ACR system to properly route calls, the ACR system generally interprets the intent of the customer, identifies the type or category of the customer call, and selects a routing destination. Speech recognition technology generally allows an ACR system to recognize natural language statements so that the caller does not have “key-in” the caller's intent. “Keying-in” data on a cell phone while driving can be a dangerous and cumbersome task. There are many different technologies that can be utilized to implement ACR systems and these technologies are often commingled on a single call. For example, a caller may first be greeted by a main ACR system that recognizes a caller's “natural language” and then transferred to an ACR sub-system that can only process touch-tone inputs. When a caller is moved between ACR systems, the change in contexts or interface characteristics can confuse the caller, reducing the call processing efficiency. Hence, there is a need for an improved automatic call routing system.